Paul J Stevens wrote:
Matthew T. O'Connor wrote:
I'll changing the unique key on headername to a unique key on
lower(headername)
OK, I thought about that, but I wasn't sure if it was legal, glad to
hear it is.
No no, don't just take my word for it. I'm not a postgres expert at all.
But at least for postgres-8.1 this works. I have *not* tested on any
other versions.
That's not how I meant it, I'm fairly sure that indexes on functions
such as lower() work fine for all the PG release that DBMail supports.
What I was getting at is, does it matter to DBMail that we lose the
difference between "Subject" and "subject"? I supposed it's preserved
in the original message body, we are only talking about the cached
headers so I suppose it would be OK, but I'm not the DBMail expert.
Matt
